The vast polar structure – VPOS – of satellite objects around the Milky Way

After the worrisome news for dark matter in the last weeks, we have to add another today, based on our research (and there is more to come very soon). We show that the disc of satellite galaxies is only a part of a bigger structure: a vast polar structure (VPOS) of diverse satellite objects surrounds the Milky Way, unexpected from cosmological models. The work was done at the University of Bonn, largely through the support of the German Research Foundation (DFG) via its priority programme 1177 “Witnesses of Cosmic History: Formation and Evolution of Black Holes, Galaxies and their Environment” and partially with support from the Bonn-Cologne Graduate School of Physics and Astronomy (BCGS).

With the increasing resolution of cosmological simulations of structure and galaxy formation, it became possible to make predictions on smaller scales. In particular, it became apparent that the dark matter subhalos, typically identified as the sites of luminous satellite galaxies around a host galaxy, are more abundant than observed galaxies. This has been termed the Missing Satellites Problem. (The name, by the way, is an interesting contortion of perspective: taking the model for granted, it blames the observed number of satellites to be too low. One could have termed it the “Over-Abundance of Subhalos Problem”, or something similar.)

But the Missing satellites problem, its suggested solutions and the problems appearing with them are not the topic of today. Instead, let’s focus on a different, more roboust test for cosmological models: the spatial distribution of subhalos / satellite galaxies. More roboust, because baryonic physics should not have an effect on scales of tens to hundreds of kpc, which are the typical observed distances between the satellites of the Milky Way.

The characteristic prediction of cold-dark-matter models is well illustrated in the following video, showing a dark matter halo similar to that assumed for the Milky Way. The ‘camera’ zooms in to the center of the halo, the size of the area shown is given in kpc the upper left corner.



It visualizes thedark matter density resulting from the Via Lactea-2 simulation (via lactea project, Jürg Diemand). The bright spots show the positions of dark matter subhalos, which might host luminous satellite galaxies. Not only does the number of predicted subhalos (> 1000) not match the observed number (currently 24, probably a few more), they are distributed rather evenly around the center, where the Milky Way would be situated. Simply put, there are subhalos in every direction.


The distribution of satellite galaxies

The MW satellites are distributed differently: they trace a disk of satellites (DoS), a planar distribution that is perpendicular to the Milky Way disc. With a radius of up to 250 kpc, this planar structure has a thickness of only 50-60 kpc.  That it is incompatible with the expectations from the standard cosmological model has been pointed out for the first time by Kroupa, Theis & Boily (2005).

It has been verified for the 11 ‘classical’ satellites (Metz et al. 2007) that they are distributed in a thin (40 kpc) planar structure that it oriented perpendicular to the Milky Way disc. This study was complemented by Metz et al (2009) with the inclusion of several fainter Galactic satellites, which led to the same orientation of the ‘disc of satellites’. Including a few additional faint satellite galaxies, we have shown in our 2010 paper (Kroupa et al. 2010) that, if you only look at the 13 faint satellite galaxies detected in the SDSS, they even independently describe the same planar orientation as the 11 classical ones.

But not only are the satellite galaxies distributed in this plane, they also move within it, as Metz et al. (2008) have shown. They looked at the 8 satellite galaxies for which the proper motion is known (that is the direction of motion on the sky, in addition to the radial velocity). Out of these 8, 7 are in agreement with moving within the plane described by all the satellites’ positions.

All of this work was done exclusively by members of the Stellar Populations and Dynamics Research (SPODYR) group at the University of Bonn.


A new idea, adding more orbit-information: streams of stars and gas

In our most recent work, we now added more different objects to this polar structure. The idea came to my mind when I (Marcel) was reading a paper about several newly discovered stellar streams (Grillmair 2009). When an object (a star cluster or dwarf galaxy) orbits around the Milky Way it looses stars. These stars will either be slightly faster or slightly slower than the object, and therefore take over or fall back along the orbit of the object. Some streams can be clearly assigned to an object, in other cases the object has been completely torn apart and only the stream is left. In both cases, the stars in the stream move more or less into the same direction as the object they came from. The streams, therefore, are situated in the same orbital plane at the progenitor object. They tell us about the object’s path around the Milky Way.

If the distribution of satellite objects around the Milky Way is stable, then the objects should move within this plane. So the orbital planes, traced by the streams of stars (or gas in some cases), should preferentially align with the satellite galaxies distributed in the DoS.

After the idea was born, we developed a method for determining the orientation of a stream. The results of the first few streams looked very promising, and so I searched the literature in order to collect data for as many streams around the Milky Way as possible. In the end, we were able to use 14 streams.

Half of them turned out to be well aligned with the DoS. If they would have been drawn from an isotropic distribution (a zeroth-order approximation of the distribution expected from cosmological models) the likelihood to find that many streams this close to the DoS is only 0.3 per cent.

Since now we know that the satellite galaxies and half of all streams align in the same structure, we began referring to this structure as the ‘vast polar structure’ (VPOS) of the Milky Way.


A different class of objects in the VPOS

The streams, as mentioned before, not only originate from dwarf galaxies, but also from globular clusters. When streams stemming from globular clusters are also located in the VPOS, shouldn’t these dense stellar systems be distributed within it, too? We checked this idea in the paper.

To do this, one has to know that globular clusters (GCs) can be classified in different groups, which are thought to have different origins (e.g. Mackey & van den Bergh 2005). There are GCs that lie in the disc of the Milky Way, so their distribution must be in the Galactic plane and not perpendicular to it like the VPOS. There are so-called old halo GCs. Their very old ages tall us that they have formed together with the Milky Way. They should not show signs of the VPOS. And there are ‘young’ halo GCs, which exhibit similarities to GCs associated with satellite galaxies, and must have a different origin than the other two groups.

When analysing the distributions of these different groups, we were stunned how well our expectations were met. The first two groups of GCs turned out to be completely unrelated to the VPOS. But the young halo globular clusters in fact define the same vast polar structure as the satellite galaxies, their motions and the streams. Even if you look only at the near young GCs (within 20 kpc) and the far ones (beyond 20 kpc) individually, they follow virtually the same structure. The likelihood of this happening in a random distribution is only 0.1 per cent.

The video below illustrated the distribution of all the different objects forming the VPOS, extending from 10 out to 250 kpc around the Milky Way. Note that the streams are magnified by a factor of three for better visibility.




The vast polar structure – VPOS – about the MW in Cartesian coordinates. The movie rotates the view over 360 degree, adding different objects around the Milky Way galaxy. The y-axis points towards the Galactic north pole. The 11 classical satellites are shown as yellow dots, the 13 new satellites are represented by the smaller green dots, young halo globular clusters are plotted as blue squares. The red curves connect the anchor points of streams of stars and gas, the (light-red) shaded regions illustrating the planes defined by these and the Galactic centre. Note that the stream coordinates are magnified by a factor of 3 to ease the comparison. The obscuration-region of 10 degree around the Milky Way disc is given by the horizontal grey areas. In the centre, the Milky Way disc orientation (edge-on) is shown by a short horizontal cyan line. One can clearly see when the view is edge-on to the VPOS: The extend of all types of objects becomes minimal, also the streams align preferentially with this structure. From standard dark matter cosmology, a much more spheroidal distribution of objects around the Milky Way is expected. We therefore propose the satellite galaxies of the Milky Way to be Tidal Dwarf Galaxies. Feel free to download the movie to be used in talks.


Suggested solutions (and why they do not work)

Within the standard cold dark matter cosmological model, such a strongly correlated structure was not predicted. This is why there have been several attempts to explain planar distributions of satellites after it became known. The list below motivates why none of these are satisfying:

  • Chance alignment: This is one of the initial and most simple ideas. If the structure is made up of only a few objects, it might just be bad luck that they all fall into a planar distribution right now. But the addition of more and more objects has reduced the likelihood of a chance alignment. A chance-alignment of the positions of the 11 classical satellites alone can be excluded at a 99.5 confidence level (Metz et al. 2007). Including the correlated motions of these or adding more objects makes this statement even more stringent, such that a chance alignment must be excluded. In addition, the preferred motion of the objects within the plane (from proper motions and stream orientations) shows that the structure is stable over time.
  • Group infall: Maybe a number of satellite galaxies were accreted by the Milky Way together in a group. It is known that associations of dwarf galaxies exist, so this was a good idea put forward by Li & Helmi (2008) and D’Onghia & Lake (2008). But as Metz et al. (2009) have shown, the observed associations are much wider than the VPOS. A structure as thin as observed can not have formed this way. In addition, the increased number of satellite objects within the VPOS speaks against this scenario, because in addition to the infall of a group of a few subhalos, a more evenly distributed population of subhalos has to be around.
  • Filamentary accretion: As seen from simulations of structure formation in the universe, there is a giant ‘cosmic web’ of material connecting galaxies, which are formed preferentially within such filaments. Maybe small dark matter halos / dwarf galaxies are accreted preferentially along such filaments, resulting in a preferred spacial distribution? This seems not to work out because the filaments, like the groups before, are too thick and not the only source of subhalos. While there are overdensities of infall-directions at large distances (see for example Libeskind et al. 2011), no structure as well defined as the VPOS is produced. Nevertheless, the abstract of Lovell et al. (2011) claims that “Quasi-planar distributions of coherently rotating satellites, such as those inferred in the Milky Way and other galaxies, arise naturally in simulations of a ΛCDM universe”. In a few days we will show in detail why this claim is unjustified. UPDATE: The preprint of our paper “Can filamentary accretion explain the orbital poles of the Milky Way satellites?” is now available on the arXiv. We will blog about it soon. UPDATE: Here is our blog post on this paper.


A radically different scenario

In our paper, we propose a radically different alternative: the VPOS has been formed from the debris of a collision of two galaxies. The satellite galaxies would then not be dark-matter dominated objects, but tidal dwarf galaxies that formed within the tidal debris stripped from another galaxy. It is noteworthy that this had already been hinted at by the early stellar-dynamical work by Kroupa (1997)  who showed that the satellite galaxies may not require dark matter but that they may appear as if they had dark matter. Tidal dwarf galaxies are observed to form and naturally align in the plane of the interaction.

As we have shown in our paper “Making counter-orbiting tidal debris. The origin of the Milky Way disc of satellites?”, Pawlowski et al. (2011), a number of features of the satellite galaxy population of the Milky Way are consistently explained if they stem from tidal debris. In addition, Pavel has laid out more reasons in favor of a tidal origin in his recent paper (Kroupa 2012).


More Information

The papers reporting problems for dark matter keep coming in more frequently lately. Be prepared for another one from our side next week, discussing filamentary accretion as a possible origin of the VPOS. UPDATE: preprint available here, discussed in the blog here.

The paper this post is based on: “The VPOS: a vast polar structure of satellite galaxies, globular clusters and streams around the Milky Way”, by Marcel S. Pawlowski, J. Pflamm-Altenburg and Pavel Kroupa. It has been accepted for publication in MNRAS and a preprint can be found on the arXiv.

The Royal Astronomical Society has published a press release on this topic, too: “Do the Milky Way’s companions spell trouble for dark matter?

It was picked up by a number of news sites, a selection of which we list below:

Blog posts:


By Pavel Kroupa and Marcel Pawlowski  (28.04.2012): “The vast polar structure – VPOS – of satellite objects around the Milky Way” on SciLogs. See the overview of topics in  The Dark Matter Crisis.

Dark Matter gone missing in many places: a crisis of modern physics?

On The Dark Matter Crisis, we have already presented numerous problems that appear within the LCDM model of cosmology. Some of these have been given names, like the “Missing Satellites Problem”, where LCDM predicts more dark matter subhaloes around the Milky Way than there are observed satellite galaxies, which are expected to trace them. Or the “Missing Baryons Problem”: from cosmological predictions we expect a certain density in the baryonic, luminous and thus in principle observable matter. But when you add up all the visible matter you observed, you only get 10-40 per cent of what you expect. The larger fraction is missing.

Even the ongoing non-detection of the DM particle in direct-detection experiments might be seen by some as another of these problems. So, there are several cases in which the model predicts something which then is not observed, thus leading to the ‘missing’ of that particular entity or observation thereof.

This week, two additional studies claim that even more seems to be missing (when your expectations are based on what LCDM predicts, that is). They both suggest a serious lack in the amount of expected dark matter on two very different size-scales: the local universe and our immediate neighborhood within the Milky Way.


Dark Matter missing in … the Local Universe

In the work titled “Missing Dark Matter in the Local Universe”, Igor D. Karachentsev has looked at a sample of 11,000 galaxies in the local Universe around the MW. He has summed up the masses of individual galaxies and galaxy-groups and used this to test a very fundamental prediction of LCDM.

The idea is as simple as it is brilliant: cosmology has precise predictions as to what is the content of our universe. In particular, it predicts the density of matter to be Ωm,glob = 0.28 +- 0.03 (83 per cent of this in dark, 17 per cent in luminous matter). Now, to test this, all you have to do is to sum up all the mass within a certain volume of space, and you can estimate the actual density of mass within that volume. To be sure that your volume is representative, it needs to be large. If you only sum over, say, a sphere of 100 kpc in diameter, the density strongly depends on whether you have a galaxy in this volume or not. Karachentsev chose to use a volume with a radius of 50 Mpc around the MW. On this size-scale, the density is expected to fluctuate by only 10 percent, a reasonably low value in astronomy. The scale can thus be assumed to be representative and you should observe the mass density predicted by LCDM.

Except that you do not.

Karachentsev reports that the average mass density is only Ωm,loc = 0.08 +- 0.02, a factor of 3-4 lower than predicted and can not be explained by the uncertainties in the data or prediction. As most of the mass-content in the Universe is supposed to be dark matter, this means that most dark matter is missing in this volume.

It is not straight-forward to interpret this result, except that it might be a serious problem for LCDM. In the paper three solutions within the framework of standard dark matter cosmology are suggested. First of all, we might resort to the unsatisfying claim that the local Universe is exceptionally non-representative of the Universe as a whole. We would then sit in a local void, a very large under-dense region of the Universe. Unfortunately, as Karachentsev states in his paper, this is in contradiction to observations. The other two suggested solutions are based on the idea that maybe not all mass is counted. Dark matter is defined to be an elusive thing, after all. Dark halos might be more extended than predicted in the models, pushing it outside the virial radius of a halo, the region in which observations can indirectly ‘measure’ it from the dynamics. However, taking this as a solution to the observed mass-deficit “clearly contradicts the existing observational data”, as Karachentsev states in his work. But maybe much of the dark matter is hiding somewhere else? Karachentsev suggests it to be in massive dark clumps not filled with galaxies (he calls them ‘dark attractors’), and thus is invisible to us when looking for galaxies only. But how could these dark clumps, with masses of galaxy-clusters, remain dark? You would need to separate the baryonic, luminous matter from a large bunch of dark matter to make sure no galaxies from in the dark attractor.

In any case, these suggestions require modifications to the behavior of dark matter because their processes are not predicted in current models. None of these possibilities seem very attractive, leaving us with the conclusion that, assuming we live in a LCDM universe, a large fraction of the dark matter is gone missing.


Dark Matter missing in … the Solar Neighborhood

The amount of dark matter in the solar neighborhood was investigated in the work “Kinematical and chemical vertical structure of the Galactic thick disk II. A lack of dark matter in the solar neighborhood” by Christian Moni Bidin and collaborators. For a short introduction, you can have a look at this proceedings paper, and yesterday, the ESO also issued a press release about this work, titled “Serious Blow to Dark Matter Theories?”.

In their work, Moni Bidin et al. have looked at a sample of 400 red giant stars close to the Sun at vertical distances of 1.5 to 4 kpc above the MW disc. In addition to the stellar 3D positions, they have derived three-dimensional kinematics for these stars. From this data, they estimate the dynamical surface mass density of the MW within this range in heights from the disc. This surface mass density should be the sum of all mass, visible and dark. But it turns out, according to their analysis, that the visible mass alone is already a perfect fit to the observed value. According to the authors, no additional mass is needed (see their plot below).

Figure 1 of Moni Bidin et al. (2012)

CAPTION: Upper panel of figure 1 of Moni Bidin et al. (2012). Observational results (black) for the surface mass density within a certain distance from the Galactic plane (x-axis). The dotted and dashed lines show the 1- and 3-sigma strip of the observations. The predictions of models (grey) containing a dark matter halo all lie significantly above the observed value, except for the model accounting for visible mass only (labelled VIS).

Their analysis is based on a number of assumptions about the structure of and kinematics in the Milky Way disc, like that the density decays exponentially in both radial and vertical direction, that there is a flat rotation curve, thar there is no bulk motion of stars in vertical or radial direction and so on. It might well be that some of their assumptions are not perfectly valid. However, they have checked that changing one of their adopted input parameters or assumptions can not solve the problem of missing DM. Very exotic hypotheses (they mention an unreasonably thin thick disc as an example) can make their data fit with the expectations from DM models, but such a solution is unsatisfying and rather improbable, according to them.

Taken together, the work suggests that, given their assumptions about the the MW disc, dark matter halos as predicted by current models do not explain the observations. It might be more informative to state it the other way around, though: according to them, the observations can be easily explained with the visible matter of the Milky Way disc alone, there is no need for more.

Note added on 21.05.2012: In a recent posting on astro-ph Bovy & Tremaine point out that the deduced amount of dark matter depends on the assumptions that go into the modelling of the stellar kinematics. They assume Newtonian dynamics to be valid (as Moni Bidin et al. have) but in contradiction to Moni Bidin et al. they show that it is not correct to assume the mean azimuthal velocity is independent of Galactocentric cylindrical radius. Instead, taking the circular velocity to be independent of the radius, Bovy & Tremaine show that the usual local matter density is arrived at. If Milgromian dynamics were correct rather than Newtonian dynamics, then it emerges that the local stellar kinematics ought to show evidence for phantom dark matter (e.g. Fig.12 in Kroupa 2012). We remind the reader that in the past it has been claimed that local stellar kinematics shows evidence for significant amounts of dark matter in the disk of the Milky Way, while more thorough later analysis has found this signal to go away (Kujiken & Gilmore 1989; Kuijken 1991Flynn & Fuchs 1994). Thus, all in all, the Newtonian analysis by Bovy & Tremaine not only “saved dark matter“, but more importantly although unintentionally, Bovy & Tremaine demonstrated consistency of the data with MOND. End Note.


Dark Matter missing in … well, it is simply not there at all

Indeed, a 50 page review of the observational tests of the standard model has been compiled by Pavel Kroupa in “The dark matter crisis: falsification of the current standard model of cosmology” and will appear in the Publications of the Astronomical Society of Australia (PASA-CSIRO publishing). Using a huge number of different data, Pavel Kroupa performs a strict logical falsification of the currently standard cosmological model, which is based on Einstein’s theory of general relativity, concluding that cold or warm dark matter cannot exist.

Note added on 21.05.2012: The implications of the Dual Dwarf Galaxy Theorem of the Kroupa 2012 paper is that cold or warm dark matter cannot be dynamically relevant in galaxies. It then implies that non-Newtonian (e.g. Milgromian) dynamics must be valid. Ironically, when interpreting Milgromian systems with Newtonian eyes, the observes will see evidence for dark matter. However, this is phanotm dark matter and it is exactly coupled to normal matter. That is, phanton dark matter is not constituted of ballistic particles which are on individual orbits within a Newtonian potential. End Note.


A crisis of modern physics

If there is no dynamically relevant cold or warm dark matter then we still need to explain the flat rotation curves of galaxies. This leads to a crisis in modern physics, as our very understanding of space-time and matter are now at stake.

Other posts you might find interesting:

II. The Fritz Zwicky Paradox and its solution

Question C.II: MOND works far too well !

Question C.III: Fundamental theoretical problems

By Pavel Kroupa and Marcel Pawlowski  (19.04.2012): “Dark Matter gone missing in many places: a crisis of modern physics?” on SciLogs. See the overview of topics in  The Dark Matter Crisis.

German TV tip: physics at the verge of collapse – science in the dark

A short TV-tip for our German readers: on April 17 (today) at 18:30 on 3sat there will be a "nano spezial" about fundamental problems of physics and cosmology, dark matter and dark energy. It is titled "Physik vor dem Kollaps – Die Wissenschaft steht im Dunkeln". It includes an interview with Pavel Kroupa.

For those without TV: the programme can already be found online in the 3sat Mediathek and will be available for the next seven days.

Question D: What about the Bullet cluster? And what about the Train-Wreck cluster Abell 520?


One result is very definite by now: neither the Bullet nor the Train Wreck clusters support (nor do they prove) the existence of cold or warm dark matter. And, they certainly do not disprove MOND. Quite on the contrary, according to current knowledge, they falsify the concordance cosmological (or LCDM) model.

The Bullet cluster consists of two clusters of galaxies that have penetrated each other leaving behind a slab of gas while the now seperating clusters retain matter as revealed through gravitational lensing. Assuming General Relativity (GR) to be valid the lensing measurements tell us that collisionless dark matter must be present in the separating clusters. But, it has been shown that the relative velocity of the two clusters need to be so large that the observed constellation ought to not occur in the real universe if it were described by GR, i.e. by the concordance cosmological model. Instead, it turns out that MOND-based models can readily account for the large relative velocity and the lensing signal as long as both clusters contain some hot dark matter or, alternatively, gas in cold clouds that cannot be detected. The Train Wreck cluster shows the opposite behaviour: assuming GR to be valid, the putative cold or warm dark matter has separated from the galaxies in this other collision of galaxy-clusters. The core of dark matter is evident from gravitational lensing (assuming GR to hold). This is inexplicable within GR because there is no known physical mechanism known for separating the dark matter from the galaxies as it does not dissipate like gas. In MOND-based models, the train wreck is also a challenge, but in principle it may perhaps be possible to separate the hot-dark-matter cluster core and the galaxies, and/or to obtain spurios lensing signals suggesting matter concentrations where there are none. Thus, the train wreck may, in the end, turn out to be a case supporting MOND-based models over GR-based ones.




As introduced in the previous contribution to The Dark Matter Crisis, Question A: Galaxies do not work in LCDM, sociology and majority views, PK had been contacted by a few people, and here are excerpts from some of the questions asked and the replies. These help to illustrate some of the issues at hand. The questions are

A) So the LCDM model fails on scales smaller than about 8 Mpc?

B1) What is a galaxy?

B2) What is a galaxy? (Addendum on the relaxation time)

C) What are the three best reasons for the failure of the LCDM model?

I: Incompatibility with observations

II: MOND works far too well !

III: Fundamental theoretical problems

D) What about the Bullet cluster?  And what about the Train-Wreck cluster Abell 520?  (this contribution)

E) Why is the main stream community so reluctant to  go along with accepting the failure of LCDM?

This contribution deals with Question D, while an upcoming contribution will concentrate on the remaining question. Beyond that we will keep posting on issues of relevance for the paradigm change.

The full question posed was “And how do you respond to the bullet cluster results which seems to point to a center of mass that does not match luminosity via weak gravitational lensing


We augment the answer with a brief discussion of the Train-Wreck cluster Abell 520. Our pevious contributions on this issue are:


A brief background to galaxy cluster dynamics:

In the Einsteinian/Newtonian theory of gravitation, galaxy clusters require about ten times as much mass in cold (or warm) dark matter than is present in normal matter.

Assuming MOND, the observed gravitational lensing and the observed kinematics in galaxy clusters merely require about a factor of two to perhaps three in additional mass. In MOND, the problem of missing mass in galaxy clusters is therefore significantly reduced. It may be completely removed if the missing mass is normal matter which is in undetectable cold gas. Or, the missing mass may be in agreement with particle physics because neutrinos oscillate and thus must have a mass. This implies the existence of additional neutral particles such as sterile neutrinos. If sterile neutrinos have a mass near 11 eV (see below) then the dark-matter problem in MOND-galaxy-clusters and in MOND-cosmological models disappears (Angus & Diaferio 2012). Such dark matter is “hot”, i.e. after the Big Bang the particles had relativistic (extremely high) velocities, and so such dark matter cannot agregate into galaxies but can be captured into galaxy-cluster-sized gravitational bodies which have sufficiently deep potential wells for the hot dark matter to not be able to escape.

This is nicely explained by Sanders (2003)  in his research paper “Clusters of galaxies with modified Newtonian dynamic., and in his review of MOND (Sanders 2009), and in the recent 160 page “Modified Newtonian Dynamics: A Review” by Famaey & McGaugh (2012).

Another approach, Modified Gravity (MOG), can deal with lensing and galaxy cluster observations entirely without dark matter (e.g. Moffat, Rahvar & Toth 2012). [note added on 16.04.2012]


Answer to the Bullet cluster:

According to Tom Shanks (private communication 2010), the data reduction to get the actual weak-lensing matter distribution map is very complex and relies on subtraction of a background. There is some freedom and it is difficult to extract a signal. See also the comment by “JR” quoted below.

But, accepting the data reduction which has been published, the Bullet cluster surprisingly turns out to be a counter-argument against the validity of the LCDM model. In LCDM the required velocities of the two clusters is too high (about 3000km/s), a velocity which does not occur. Now, in a MONDian universe, such velocities occur rather naturally, and so with some hot dark matter (in fact the same as I wrote above, 11eV particles) one gets beautiful agreement with the observations!

The Bullet Cluster (1E 0657-56) is often perceived to be a disproof of Milgromian dynamics because even in Milgromian dynamics DM is required to explain the observed separation of the weak lensing signal and the baryonic matter. In actuality, the Bullet Cluster is, if anything, a major problem for the LCDM model because the large relative cluster–cluster velocity at the mass scale of the two observed clusters required to provide the observed gas shock front cannot be attained in the LCDM model, as shown by Lee & Komatsu (2010) in their research paper “Bullet Cluster: A Challenge to ΛCDM Cosmology” and as verified and deepened by Thompson & Nagamine (2012) in their research paper “Pairwise velocities of dark matter haloes: a test for the Λ cold dark matter model using the bullet cluster“. Thomposn & Nagamine  “conclude that either 1E 0657-56 is incompatible with the concordance ΛCDM universe or the initial conditions suggested by the non-cosmological simulations must be revised to give a lower value of” the relative velocity.

But the high relative velocities between the two sub-clusters in the Bullet cluster arise naturally and abundantly in a Milgromian cosmology:

Assuming the Milgromian framework to be the correct description of effective gravitational dynamics, it has been shown that the Bullet Cluster lensing signal can be accounted for in it. Ibn their researhc paper,  “Can MOND take a bullet? Analytical comparisons of three versions of MOND beyond spherical symmetry“,  Angus, Famaey & Zhao (2006) state “In particular, we can generate a multicentred baryonic system with a weak lensing signal resembling that of the merging galaxy cluster 1E 0657-56 with a bullet-like light distribution.

If a Milgromian cosmology is allowed to have a hot DM component then the Bullet Cluster is indeed well explainable. In the research paper on “The collision velocity of the bullet cluster in conventional and modified dynamics“, Angus & McGaugh (2008) they summarise:

“We consider the orbit of the bullet cluster 1E 0657-56 in both cold dark matter (CDM) and Modified Newtonian Dynamics (MOND) using accurate mass models appropriate to each case in order to ascertain the maximum plausible collision velocity. Impact velocities consistent with the shock velocity (~ 4700kms-1) occur naturally in MOND. CDM can generate collision velocities of at most ~3800kms-1, and is only consistent with the data, provided that the shock velocity has been substantially enhanced by hydrodynamical effects. “

Using a new cosmological N-body code for MOND, Angus & Diaferio (2011) find “As a last test, we computed the relative velocity between pairs of haloes within 10 Mpc and find that pairs with velocities larger than 3000 km s-1, like the bullet cluster, can form without difficulty.

We know that neutrinos oscillate, therefore they must have a mass. That mass is small. This makes them a form of hot DM that we most definitely know to exist. In order to explain the oscillations, particle physics suggests the possible existence of more massive, sterile neutrinos, which interact by gravity. If they exist they might be massive enough to account for the missing mass in galaxy clusters in MOND (and they can fit the first three acoustic peaks in the CMB). A research paper discussing the possible role of sterile neutrinos for dark matter has been published by Dodelson & Wildrow (1994).

Taking this ansatz, Angus, Famaey & Diaferio (2010) demonstrate, in their research paper “Equilibrium configurations of 11 eV sterile neutrinos in MONDian galaxy clusters”  that consistency in solving the mass-deficit in galaxy clusters and accounting for the CMB radiation power spectrum is achieved if sterile neutrinos (SN) have a mass near 11 eV. They write “we conclude that it is intriguing that the minimum mass of SN particle that can match the CMB is the same as the minimum mass found here to be consistent with equilibrium configurations of Milgromian clusters of galaxies


The Train Wreck cluster:

The Train-Wreck cluster (Abell 520) has been shown to be incompatible with the LCDM model because the putative C/WDM particles have separated from the galaxies such that a core of DM is left behind and away from the concentrations of galaxies, as Mahdavi et al. (2007) find in their research paper “A Dark Core in Abell 520“. There is no known physical mechanism which can separate cold or warm dark matter from galaxies to the extend required by the Train Wreck.

Jee et al. (2012) return to the Train Wreck with their research paper “A Study of the Dark Core in A520 with the Hubble Space Telescope: The Mystery Deepens“, confirming the problem. They speculate on a possible solution such as DM possibly having a self-interaction property, and interestingly they avoid discussion of any alternative theory of gravitation.

The Train Wreck remains not understood.

As pointed out by Kroupa (2012), in a MONDian cosmological model with hot dark matter (HDM) it is conceivable, at least in principle under certain conditions, for the self-bound galaxy-cluster-sized HDM core of the whole cluster to dissociate itself from the baryonic matter in galaxies since the galaxies do not reside in HDM halos. Each individual galaxy would remain on the baryonic Tully-Fisher relation, as is observed to be the case for all disk galaxies and as is required to be the case if MOND is correct (e.g. Famaey & McGaugh 2012). And, in MOND-based models it may perhaps be possible to obtain spurios lensing signals suggesting matter concentrations where there are none.

Finally, in the comment “Scientific Polemicism” to our previous contribution  “The Train Wreck Cluster – an “anti-Bullet-Cluster”: disproof of Cold or Warm Dark Matter, “JR” writes on the 18.10.2010 at 14:45:

The main reason why most scientists remain sceptical about the Abell 520 “train wreck” results is that different groups analysing the *same data* obtain different mass maps (see Okabe & Umetsu 2008). Now that’s a train wreck! The same cannot be said for the bullet cluster, where – to the best of my knowledge – all authors currently agree on the lensing mass maps. This does not mean, of course, that the bullet is right and Abell 520 is wrong – we should remain open minded about both. But I am particularly sceptical of the Abell 520 results because of a well-known problem with lensing mass reconstruction: the monopole degeneracy. This was illustrated beautifully in recent work by Liesenborgs et al. (2008) who show that the monopole degeneracy can lead to phantom peaks in the mass distribution (see their Figure 3). Their work focused on strong rather than weak lensing, but weak lensing suffers from exactly the same problems.


Within the modified gravity (MOG) framework, Moffat & Toth (2009) and Moffat, Rahvar & Toth (2012) argue to be able to account for both the Bullet and the Train Wreck cluster.


A ring of dark matter?:

Milgrom & Sanders (2008) analyse in their research paper “Rings and Shells of “Dark Matter” as MOND Artifacts” the recent detection using weak lensing of a ring of dark matter around a galaxy cluster. They write in their abstract:

We consider the possibility that this pure MOND phenomenon is in the basis of the recent finding of such a ring in the galaxy cluster Cl 0024+17 by Jee et al. (2007). We find that the parameters of the observed ring can be naturally explained in this way; this feature may therefore turn out to be direct evidence for MOND.


By Pavel Kroupa and Marcel Pawlowski  (15.04.2012): “Question D: What about the Bullet cluster? And what about the Train Wreck cluster Abell 520” on SciLogs. See the overview of topics in  The Dark Matter Crisis.